Use AI to find loved ones lost after the Holocaust!

New business opened up in the field of AI facial recognition?

This time, it is to identify face images in old photos during World War II.

Recently, Daniel Patt, a software engineer from Google, developed an AI face recognition technology called N2N (Numbers to Names), which can identify photos of Europe before World War II and the Holocaust period, and compare them with modern people. Get in touch.

 

Using AI to find long-lost relatives
In 2016, Pat had an idea while visiting the Polish Jewish Memorial in Warsaw.

Are these unfamiliar faces related to him by blood?

Three of his grandparents/grandparents were Holocaust survivors from Poland, and he wanted to help his grandmother find pictures of family members killed by the Nazis.

During World War II, due to the large number of Polish Jews who were detained in different concentration camps, many of them were unaccounted for.

Just through a yellowed photo, it is difficult to identify who the face is, let alone find his lost relatives.

 

So, he returned home and immediately turned this idea into reality.

The software was originally conceived to collect image information of faces through a database and use artificial intelligence algorithms to help match the top ten options with the highest similarity.

Most of the image data comes from The US Holocaust Memorial Museum, in addition to more than a million images from databases across the country.

Users only need to select an image in the computer file, click upload, and the system will automatically filter out the top ten options with the highest matching images.

 

In addition, users can also click the source address to view the year, location, collection and other information of the picture.

picture

One drawback is that if you input modern images of people, the search results can also be outrageous.

 

 

 

In short, the system functions still need to be improved.

In addition, Patt works with other teams of software engineers and data scientists at Google to improve the scope and accuracy of searches.

Due to the risk of privacy leakage in the face recognition system, Patt said, “We do not make any evaluations of identities. We are only responsible for presenting the results with similarity scores and letting users judge for themselves.”

 

So how does this technology recognize faces?

Initially, face recognition technology had to start with “how to judge that the detected image is a face”.

In 2001, computer vision researchers Paul Viola and Michael Jones proposed a framework to detect human faces in real time with high accuracy.

This framework can understand “what is a face and what is not” based on the training model.

After training, the model extracts specific features, which are then stored in a file so that the features in the new image can be compared at various stages with previously stored features.

To help ensure accuracy, the algorithm needs to be trained on “a large dataset of hundreds of thousands of positive and negative images,” improving the algorithm’s ability to determine whether there are faces and where they are in the images.

If the image under study passes each stage of feature comparison, the face has been detected and the operation can continue.

Although the Viola-Jones framework is highly accurate for recognizing faces in real-time applications, it has certain limitations.

For example, the framework may not work if the face is wearing a mask, or if a face is not properly oriented.

To help eliminate the shortcomings of the Viola-Jones framework and improve face detection, other algorithms have been developed.

Such as region-based convolutional neural network (R-CNN) and single-shot detector (SSD) to help improve the process.

A Convolutional Neural Network (CNN) is an artificial neural network for image recognition and processing, specifically for processing pixel data.

R-CNN generates region proposals on the CNN framework to localize and classify objects in images.

While methods based on region proposal networks, such as R-CNN, require two shots—one for generating region proposals and another for detecting objects for each proposal—SSD requires only one shot to detect multiple objects in an image. object. Therefore, SSD is significantly faster than R-CNN.

In recent years, face recognition technology driven by deep learning models has significant advantages over traditional computer vision methods.

Early face recognition mostly used traditional machine learning algorithms, and the research focus was more on how to extract more discriminative features and how to align faces more effectively.

With the deepening of research, the performance improvement of traditional machine learning algorithm face recognition on two-dimensional images has gradually reached a bottleneck.

People began to study the problem of face recognition in video, or combined the method of 3D model to further improve the performance of face recognition, while a few scholars began to study the problem of 3D face recognition.

 

On the most famous LFW public library, the deep learning algorithm broke through the bottleneck of the traditional machine learning algorithm’s face recognition performance on two-dimensional images, and increased the recognition rate to more than 97% for the first time.

That is, the “high-dimensional model established by CNN network” is used to directly extract effective discriminative features from the input face image, and directly calculate the cosine distance for face recognition.

Face detection has evolved from basic computer vision techniques to advances in machine learning (ML) to increasingly complex artificial neural networks (ANNs) and related techniques, resulting in continuous performance improvements.

Now, it plays an important role as the first step in many key applications – including face tracking, facial analysis and facial recognition.

During World War II, China also suffered from the trauma of war, and many of the people in the photos at the time were long ago indistinguishable.

There are many relatives and friends who have been traumatized by the war from the grandparents’ generation, whose whereabouts are unknown.

The development of this technology may help people unravel the dusty years and find some solace for people in the past.